通信领域中嵌入式操作系统平台相关算法研究
摘要
随着通讯系统的发展，特别是以第三代为代表的全分组网络的发展，对目前的通讯系统软件提出了更高的要求，而过去所使用的嵌入式操作系统在很多方面已经不能满足现有的需求，需要重新设计通讯系统的运行支撑环境。
目前通讯系统普遍采用的是嵌入式软件设计，由于嵌入软件系统与通讯应用软件之间是紧耦合系统，操作系统与上层应用软件共用同一个代码和地址空间，而现在通讯系统的软件越来越庞大，设计越来越复杂，导致错误跟踪定位异常困难，特别是现场情况下的死机和跑飞常常无法追踪其原因。
在第三代核心网络中，由于系统对效率的要求非常高，系统中有大量的模块间消息传递，过去的传统交换机只需要处理通讯信令，而第三代设备中不仅通讯信令复杂了许多，还要处理用户面的数据，而用户面的数据量比信令高几个数量级，大通讯量必然导致大量的模块间消息发送以及大量的使用定时器和内存操作，目前的商用嵌入系统已经不能满足当前通讯技术发展的要求。
另外，当前通讯对实时性要求也比过去高了很多，由于将来的核心网络是全分组交换网络，如何在分组网络上保证服务质量成了一个难题，首先是要有足够的带宽，其次必须有良好的算法来保证，而这都不是目前的商用平台所能提供的。
本论文完成了在商用平台上构建一个高效的运行支撑平台，重新设计了其中的一些关键算法，为将来的通讯软件提供更加强大的功能和更高的效率。
第一章主要说明本篇论文的整体结构框架。在第二章中对嵌入式实时系统、基于VxWorks的嵌入式统一平台支撑子系统的基本知识做了详细的介绍。针对通常内存管理算法缺少内存的保护措施和
错误诊断方法，也很少有内存碎片整理机制，在第三章中主要研究了通信系统中高效内存算法并进行对比测试。第四章通过对多队列算法的不足分析，设计了单循环队列定时器算法并和多队列算法比较得出测试结果。第五章深入研究了嵌入式实时操作系统中零拷贝消息队列算法的设计和实现。
最后在第六章对本文中提出的算法在实现方面进行了全面的总结，并对嵌入式实时操作系统的发展做了一点展望。
With the development of telecom system, especially with the booming of all-packet networks represented by 3G, current telecom software is required to provide more powerful capability. However, former embedded operating systems are becoming lack of capability to meet the demand of newly developed telecom system. It is highly pressing to re-design a novel operating support platform for the telecom system.
The designs of current telecom systems generally adopt embedded software architecture. The operating system and upper application system share the same code and address space due to the tight coupling between them. However, software of telecom system is becoming more and more complicated, which makes it exceptionally difficult to trace the exceptions of programs. It is even harder to locate the exceptions on locale.
In the core network of 3G, telecom system is required to provide higher efficiency for the need of volume message transferring among modules. Contrasted with the traditional switches, equipments of 3G dispose not only much more complicated signaling, but data of user-side as well. Volume message transferring among modules definitely results in volume handling of timer and memory resources.
Besides, the future core telecom networks are All-packet switch networks, which raise higher demand for real-time ability. In order to guarantee QoS in the packet network, abundant bandwidth and well algorithms should be provided.
But these two capabilities cannot be fully provided in current commercial platforms.
This essay is to construct a highly effective operating support platform, and to re-design some of the key algorithms based on the current commercial operating system, in order to support more powerful functions and higher effectiveness for the future telecom software.
Chapter 1 describes the outline of this essay. Chapter 2 focuses on the fundamental knowledge of RTOS and Common Support Platform based on VxWorks. Because of shortage of memory protection measures and error diagnosis methods in common memory management algorithm, and few packing up memory fragmentations mechanism, the description of a novel memory management algorithm is primarily presented in chapter 3. The test result of this algorithm is also described. In chapter 4 the timer management algorithms of single circular queue is designed by analysis of shortcomings of multiple queues. The comparison between them and the test outcome are also described in this chapter. Chapter 5 presents the message queue of zero-copy in RTOS.
The last chapter of this essay gives some brief summaries on how to realize the mentioned algorithms in telecom system. The author also prospects the development trend for the RTOS at the end of this essay.
目前通讯系统普遍采用的是嵌入式软件设计，由于嵌入软件系统与通讯应用软件之间是紧耦合系统，操作系统与上层应用软件共用同一个代码和地址空间，而现在通讯系统的软件越来越庞大，设计越来越复杂，导致错误跟踪定位异常困难，特别是现场情况下的死机和跑飞常常无法追踪其原因。
在第三代核心网络中，由于系统对效率的要求非常高，系统中有大量的模块间消息传递，过去的传统交换机只需要处理通讯信令，而第三代设备中不仅通讯信令复杂了许多，还要处理用户面的数据，而用户面的数据量比信令高几个数量级，大通讯量必然导致大量的模块间消息发送以及大量的使用定时器和内存操作，目前的商用嵌入系统已经不能满足当前通讯技术发展的要求。
另外，当前通讯对实时性要求也比过去高了很多，由于将来的核心网络是全分组交换网络，如何在分组网络上保证服务质量成了一个难题，首先是要有足够的带宽，其次必须有良好的算法来保证，而这都不是目前的商用平台所能提供的。
本论文完成了在商用平台上构建一个高效的运行支撑平台，重新设计了其中的一些关键算法，为将来的通讯软件提供更加强大的功能和更高的效率。
第一章主要说明本篇论文的整体结构框架。在第二章中对嵌入式实时系统、基于VxWorks的嵌入式统一平台支撑子系统的基本知识做了详细的介绍。针对通常内存管理算法缺少内存的保护措施和
错误诊断方法，也很少有内存碎片整理机制，在第三章中主要研究了通信系统中高效内存算法并进行对比测试。第四章通过对多队列算法的不足分析，设计了单循环队列定时器算法并和多队列算法比较得出测试结果。第五章深入研究了嵌入式实时操作系统中零拷贝消息队列算法的设计和实现。
最后在第六章对本文中提出的算法在实现方面进行了全面的总结，并对嵌入式实时操作系统的发展做了一点展望。
With the development of telecom system, especially with the booming of all-packet networks represented by 3G, current telecom software is required to provide more powerful capability. However, former embedded operating systems are becoming lack of capability to meet the demand of newly developed telecom system. It is highly pressing to re-design a novel operating support platform for the telecom system.
The designs of current telecom systems generally adopt embedded software architecture. The operating system and upper application system share the same code and address space due to the tight coupling between them. However, software of telecom system is becoming more and more complicated, which makes it exceptionally difficult to trace the exceptions of programs. It is even harder to locate the exceptions on locale.
In the core network of 3G, telecom system is required to provide higher efficiency for the need of volume message transferring among modules. Contrasted with the traditional switches, equipments of 3G dispose not only much more complicated signaling, but data of user-side as well. Volume message transferring among modules definitely results in volume handling of timer and memory resources.
Besides, the future core telecom networks are All-packet switch networks, which raise higher demand for real-time ability. In order to guarantee QoS in the packet network, abundant bandwidth and well algorithms should be provided.
But these two capabilities cannot be fully provided in current commercial platforms.
This essay is to construct a highly effective operating support platform, and to re-design some of the key algorithms based on the current commercial operating system, in order to support more powerful functions and higher effectiveness for the future telecom software.
Chapter 1 describes the outline of this essay. Chapter 2 focuses on the fundamental knowledge of RTOS and Common Support Platform based on VxWorks. Because of shortage of memory protection measures and error diagnosis methods in common memory management algorithm, and few packing up memory fragmentations mechanism, the description of a novel memory management algorithm is primarily presented in chapter 3. The test result of this algorithm is also described. In chapter 4 the timer management algorithms of single circular queue is designed by analysis of shortcomings of multiple queues. The comparison between them and the test outcome are also described in this chapter. Chapter 5 presents the message queue of zero-copy in RTOS.
The last chapter of this essay gives some brief summaries on how to realize the mentioned algorithms in telecom system. The author also prospects the development trend for the RTOS at the end of this essay.
引文
[1] Jean j.Labrosse. μc/os - II 源码公开的实时嵌入式操作系统. 北京：中国电力出版社，2001
[2] Galli，D. Distributed Operating Systems: Concepts and Practice. Upper Saddle River, NJ: Prentice Hall,2000
[3] VxWorks Programmer’s Guide，WindRiver Inc, 2000
[4] VxWorks Network Programmer's Guide，WindRiver Inc, 2000
[5] William S. Operating Systems : Internals and Design Principles ( Fourth Edition ) .北京：电子工业出版社，2001
[6] 孔祥营，柏桂枝.嵌入式实时操作系统VxWorks及其开发环境Tornado.北京：中国电力出版社，2002
[7] Douglass,Bruce Powel Real-Time UML:Efficient Objects for Embedded Systems,Reading,MA: Addison-Wesley-Longman, 1998
[8] Douglass,Bruce Powel Doing Hard Time:Debeloping real-Time Systems with UML,Objects,frameworks and Patterns Reading,MA:Addison-Wesley-Longman,1999
[9] Laplante，Phillip A. Real-Time Systems Design and Analysis，An Engineer's Handbook. Piscataway，New Jersey[c]: IEEE Computer Society Press，1992
[10]Adve,S,V.,and Hill,M,D. Weak ordering-A new definition.In Proceedings of the International Sumposium on Computer Architecture (May 1990),2-14.
[11]Kelvin D.Nilsen. "Relialbe Real-Time Garbage Collection of C++". 1993 OOPSLA workshop on Garbage Collection in Object-Oriented Systems
[12]Kelvin D.Nilsen. "Cost-Effective Hardware-Assisted Real-Time Garbage Collection". The 1992 ACM SIGPLAN Conference on Programming Language Design and Implementation
[13]Kelvin D.Nilsen. "A High-Performance Hardware-Assisted Real-Time Garbage Collection". The 1993 Journal of
Programming Language Design and Implementation"
[14]Damien Doligez and Xavier Leroy. "A concurrent, generational garbage collector for a multithreaded implementation of ML". ACM-20 th POPL-1/93-S.C.
[15]Huber, HGM. "List-Fortran, a Basic List-Processing Extension of Fortran on the IBM 360". Naval Weapons Lab Dahlgren Va. 5 Oct 1967
[2] Galli，D. Distributed Operating Systems: Concepts and Practice. Upper Saddle River, NJ: Prentice Hall,2000
[3] VxWorks Programmer’s Guide，WindRiver Inc, 2000
[4] VxWorks Network Programmer's Guide，WindRiver Inc, 2000
[5] William S. Operating Systems : Internals and Design Principles ( Fourth Edition ) .北京：电子工业出版社，2001
[6] 孔祥营，柏桂枝.嵌入式实时操作系统VxWorks及其开发环境Tornado.北京：中国电力出版社，2002
[7] Douglass,Bruce Powel Real-Time UML:Efficient Objects for Embedded Systems,Reading,MA: Addison-Wesley-Longman, 1998
[8] Douglass,Bruce Powel Doing Hard Time:Debeloping real-Time Systems with UML,Objects,frameworks and Patterns Reading,MA:Addison-Wesley-Longman,1999
[9] Laplante，Phillip A. Real-Time Systems Design and Analysis，An Engineer's Handbook. Piscataway，New Jersey[c]: IEEE Computer Society Press，1992
[10]Adve,S,V.,and Hill,M,D. Weak ordering-A new definition.In Proceedings of the International Sumposium on Computer Architecture (May 1990),2-14.
[11]Kelvin D.Nilsen. "Relialbe Real-Time Garbage Collection of C++". 1993 OOPSLA workshop on Garbage Collection in Object-Oriented Systems
[12]Kelvin D.Nilsen. "Cost-Effective Hardware-Assisted Real-Time Garbage Collection". The 1992 ACM SIGPLAN Conference on Programming Language Design and Implementation
[13]Kelvin D.Nilsen. "A High-Performance Hardware-Assisted Real-Time Garbage Collection". The 1993 Journal of
Programming Language Design and Implementation"
[14]Damien Doligez and Xavier Leroy. "A concurrent, generational garbage collector for a multithreaded implementation of ML". ACM-20 th POPL-1/93-S.C.
[15]Huber, HGM. "List-Fortran, a Basic List-Processing Extension of Fortran on the IBM 360". Naval Weapons Lab Dahlgren Va. 5 Oct 1967
© 2004-2018 中国地质图书馆版权所有 京ICP备05064691号 京公网安备11010802017129号 地址：北京市海淀区学院路29号 邮编：100083 电话：办公室：(+86 10)66554848；文献借阅、咨询服务、科技查新：66554700 |